Towards a Machine Ethics

Am 15. März 2013 trägt Oliver Bendel bei der European TA Conference in Prag zum Thema „Towards a Machine Ethics“ vor. Das Abstract, das im Frühjahr im Tagungsband veröffentlicht wird, hebt mit den Worten an: „There is an increasing use of autonomous machines such as agents, chatbots, algorithmic trading computers, robots of different stripes and unmanned ground or air vehicles. They populate the modern world like legendary figures and artificial creatures in Greek mythology – with the main difference being that they are real in the narrow sense of the word. Some are only partially autonomous (acting under human command) while others are completely autonomous within their area of action. A genuinely autonomous machine should be able to act in a moral way, able to make decisions that are good for humans, animals and the environment. But what does it mean for machines to behave morally? Should they learn moral rules? Should they evaluate the consequences of their acts? Or should they become a virtuous character, following Aristotle? How is it possible to implement the classical normative models of ethics and is there a need for new ones?“ Der erste, zweite und dritte Abschnitt des Beitrags wird knapp zusammengefasst; dann wird auf den vierten eingegangen: „Fourthly, the paper tries to answer the question if and how it is possible to implement the classical normative models of ethics and which models should be preferred. Seven important normative approaches are described and estimated relating to their suitability for machine processing. Then the focus shifts to duty-based ethics, ethics of responsibility and virtue ethics that seem to be serious candidates. With a short technical analysis it can be shown that they fit to machine processing, apart from some limitations. The most promising approach may be the combination of the selected normative models. It is not only similar in the ’normal‘ human ethics, but also an opportunity to balance out weaknesses of the autonomous machines and to allow them alternatives. In addition, other methods like orientation on reference persons and social media evaluation could be used.“ Am Ende werden Wichtigkeit und Schwierigkeit des jungen Forschungsgebiets der Maschinenethik gleichermaßen betont und persönliche und allgemeine Einschätzungen gegeben.